J. AMER. Soc. HORT. SCI. 117(2):308-312. 1992.
Allelopathic Potential of Celery Residues on
Dorm G. Shilling
Agronomy Department, University of Florida,’ Gainesville, FL 32611
Joan A. Dusky
Everglades Research and Education Center, P. O. Box 8003, Belle Glade, FL 33430
Mark A. Mossier
Agronomy Department, University of Florida, Gainesville, FL 32611
Thomas A. Bewick
Vegetable Crops Department, University of Florida, Gainesville, FL 32611
Additional index words. Lactuca sativa, Apium graveolens, allelopathy, activated carbon, phytotoxins, crop residues
Abstract. Poor emergence of commercially grown lettuce has been observed when planted immediately after the
removal of a celery crop. Greenhouse experiments were conducted to evaluate the possible allelopathic effects of
celery residue on the emergence and growth of lettuce. The influence of amount and type of celery tissue, growth
medium and fertility, incubation time in soil, and amendment of growth medium containing celery residue with
activated charcoal was evaluated with respect to the allelopathic potential of celery. Celery root tissue was 1.8 and
1.6 times more toxic to lettuce seedling growth than was celery petiole or lamina tissue, respectively. Lettuce shoot
growth was inhibited to a greater extent when grown in sand amended with celery residue rather than either amended
vermiculite or potting soil. Incubation of celery root residue in soil for 4 weeks increased phytotoxicity at 1% (v/v)
and decreased it at 4% (v/v). Increasing the fertility of pure sand with varying amounts of Hoagland’s solution did
not reverse the allelopathic effects of celery residue. The addition of activated carbon to the medium increased the
growth of lettuce exposed to celery residues. Celery residues possess allelopathic potential to developing lettuce
seedlings. Celery tissue type and concentration, soil type, incubation of celery root residue in soil, and addition of
activated carbon to the growing medium influenced the magnitude of the observed phytotoxicity.
The term allelopathy has been defined by Rice (1974) as “the treated celery, suggesting that psoralen production can be stim-
direct or indirect harmful effect by one plant (including micro- ulated by Cu+2 in the absence of pathogenic organisms (Beier
organisms) on another through the production of chemical com- and Oertli, 1983). Compounds containing copper are recom-
pounds that escape into the environment.” Currently, allelopathy mended to control a number of bacterial and fungal diseases of
is understood as the inhibitory and/or stimulator effects of one celery (Sherf and MacNab, 1986).
plant (either microbial or higher plant) on another by the pro- Over a period of several years, vegetable growers in the Ev-
duction of a chemical substance that is released into the envi- erglades Agricultural Area of Florida have reported that lettuce,
ronment (Putnam and Tang, 1986). Recent reviews on the topic planted soon after the harvest of celery, emerged poorly and
are available (Duke, 1986; Putnam, 1988; Rice, 1984) and in- more slowly than normally expected (D. Botts, personal com-
clude discussion of crop residues that possess allelopathic ac- munication). The resulting lettuce stand, combined with a non-
tivity under certain environmental conditions. uniform harvest period, was commercially unacceptable. Other
Many volatile constituents of celery have been isolated crop residues, including barley (Hordeum vulgare L.) and rye
(MacLeod et al., 1988), including p- cymene, limonene, and ß- (Secale ceveale L.), reduced emergence and growth of lettuce
selinene, and are reported to have allelopathic activity (Fischer, in California (Patrick et al., 1963). The purpose of our studies
1986). Celery tissue also contains linear furocoumarins that act was to determine whether celery residue incorporated into soil
as phytoalexins against such pathogens as Sclerotinia sclero- reduces the emergence and growth of lettuce under controlled
tiorum (LIB. ) de Barry (Beier and Oertli, 1983), Erwinia car- greenhouse conditions and, if so, to study several variables that
otovora pv. carotovora (Et.) (Surico et al., 1987), and certain might influence this effect.
viruses (Lord et al., 1988). One of these compounds, psoralen,
is a potent inhibitor of seed germination (Putnam, 1988). Early Materials and Methods
research indicated that psoralen reduced Lactuca germination
and sprout growth at a concentration of 3.6 x 10–4 M (Brown, Growth media and type and amount of celery tissue. Treat-
1981). Psoralen isolated from Psoralea subacaulis Pursh. in- ment variables were arranged factorially (3 x 3 x 6) to eval-
hibited germination of lettuce seed at concentrations as low as uate any interactive effects. Three growth media were used (all
1 ppb (Baskin et al., 1967). Celery contains psoralen concen- per 425-cm 3 pot): 1) 110 g of a high organic matter potting
trations an order of magnitude higher than that needed to inhibit medium that contained 30% sphagnum peat, 50% vermiculite,
lettuce germination (Beier et al., 1983). Psoralen was identified 18% perlite, and 2% sand (by volume) (Metro-mix, Gracewood
as a major linear furocoumarin constituent of copper sulfate- Horticultural Products, Cambridge, Mass.); 2) 200 g of 100%
vermiculite; and 3) 610 g of 100% quartz sand. The media were
not sterilized before use. Celery, grown using standard com-
Received for publication 9 May 1991. Accepted for publication 28 Oct. 1991. mercial practices and obtained from the Everglades Agricultural
Florida Agr. Expt. Sta. Bul. no. R-01579. The cost of publishing this paper
was defrayed in part by the payment of page charges. Under postal regulations, Area, Belle Glade, Fla., was harvested at a time corresponding
this paper therefore must be hereby marked advertisement solely to indicate this to commercial harvest, air-dried, and stored at 0C until use.
fact. Celery tissues evaluated for allelopathic potential were divided
308 J. Amer. Soc. Hort. Sci. 117(2):308-312. 1992.
into petioles, laminas, and roots. Celery residue was incorpo- the appropriate amounts of celery residue and activated carbon
rated into the media at 0, 1.1, 2.2, 4.4, 8.8, or 13.2 g/425-cm3 were homogenized with sand, 425-cm3 pots were filled with the
pot, which corresponded to 0, 0.5%, 1%, 2%, 4%, and 6% mixtures. Lettuce was seeded 0.5 cm deep (30 seeds per pot)
(v/v). Because the media had vastly different densities, residue and the pots saturated with water by subirrigation. Lettuce was
percentages were established on a volume basis so that the per- fertilized twice a week with 10 ml of 5 x Hoagland’s solution.
centage of residues per gram of medium would be the same. Plant maintenance, harvest, and statistical design and analy-
Celery residues were incorporated into the soils by vigorously sis. Lettuce was maintained in a glass greenhouse (30 ± 5C).
shaking the components in paper bags. Immediately after in- Routine watering was accomplished by subirrigation. Nutrient
corporation, pots were filled and 30 ‘Iceberg’ lettuce (commer- solution was applied as a drench when used. Lettuce emergence
cial crisphead lettuce type) seeds per pot were planted 0.5 cm was determined 7, 14, and 28 days after the initiation of all
deep. The soil mixture was then saturated with water by subir- experiments. After 28 days of growth, lettuce was harvested,
rigation. Preliminary studies were conducted to determine let- shoot and root tissues were separated, dried at 90C for 3 days,
tuce growth in the three soil types in response to fertility (data and weighed. Accurate root weights could not be determined in
not shown). Ten milliliters of a 2 × Hoagland solution (Hoag- either potting medium or vermiculite because the media particles
land and Arnon, 1950) was applied one, two, or three times per adhered tightly to lettuce roots. No supplemental lighting was
week to lettuce growing in potting soil, vermiculite, or sand, used.
respectively. These levels of fertility input for each growth me- All treatments were replicated three times and each study
dium were shown to optimize lettuce growth (data not shown). conducted twice using a randomized complete-block design. Data.
Incubation of celery residue in soil. Treatments included three were initially analyzed using analysis of variance to test for
celery root residue percentages [0%, 1%, and 4% (v/v); 0, 2.2, treatment effects and interactions. Because there were no ex-
and 8.8 g/pot, respectively] and four incubation periods (0, 4, periment x treatment interactions (P > 0.05), data were com-
8, and 12 weeks) arranged factorially. Celery root residue was bined across experiments. Regression analysis was used to
thoroughly mixed with 370 g of an Arredondo fine sand soil determine the concentration of celery residue required to cause
(loamy, siliceous, hyperthermic Grossarenic Paleudult) as de- 50% inhibition of lettuce growth (I50 value); 95% confidence
scribed previously. The soil was kept moist throughout the ex- intervals were calculated for each I50 value (Draper and Smith,
periment by subirrigation. The 0 incubation period was established 1981). Specific information dealing with the mean separation
by planting 30 ‘Iceberg’ seeds 0.5 cm deep immediately after procedures is presented with the data.
mixing the celery residue and soil and filling 425-cm3 pots with
this mixture. Four, eight, and 12 weeks after the soil/residue Results and Discussion
mixture had been added to the pots (i.e., the different incubation Growth media and type and amount of celery tissue. Growth
periods), one-fourth of the experiment was seeded to lettuce as medium influenced the degree of response of lettuce to celery
described previously. All treatments were fertilized with 20 ml residues (Table 1). As the I50 value decreased, the phytotoxicity
of 1 × Hoagland solution at the time of planting. Lettuce emer- of the residue increased because less residue was required to
gence data were recorded at 7, 14, and 28 days after sowing, cause the same amount of growth inhibition. In terms of shoot
and lettuce shoots were harvested at 28 days after sowing. biomass, celery residue was more phytotoxic in sand than either
Fertility study. Treatments were 0, 20, 40, 50, 100,200, and vermiculite or potting medium, possibly because of the greater
400 ml of 1 × Hoagland solution applied per week both with availability of putative allelopathic compounds due to reduced
and without 2.25% (v/v) celery root residue. Lettuce was seeded adsorption by the inert sand. When lettuce was grown in ver-
0.5 cm deep (30 seeds per 425-cm3 pot) in pure quartz sand miculite or potting medium it was initially less susceptible (larger
that was saturated with water by subirrigation. Subsequent ir- I 50 values) to celery residue as indicated by emergence data.
rigation was accomplished with a combination of nutrient so- There are several possible explanations for the transient effect
lution and subirrigation. of celery residue in these media on lettuce emergence. First,
Activated carbon study. Treatments included three percent- over the 28-day study period, the toxicity of the residue could
ages of activated carbon [0%, 6%, and 12% (v/v); 0, 6, and 12 have increased due to activation and/or release of phytotoxic
g/pot, respectively] and four percentages of celery root residue compound(s). Second, some of the lettuce seedlings that ini-
[0%, 0.5%, 1.0%, and 2.0% (v/v)] arranged factorially. After tially emerged died either due to damping-off or due to some
J. Amer. Soc. Hort. Sci. 117(2):308-312. 1992. 309
direct effect from the celery residue that resulted in eventual
death (e.g., reduced root growth that would not effect initial
emergence but would influence long-term viability) (Patrick et more toxic than celery petiole and lamina tissue, respectively.
al., 1964). Third, the compound(s) active in celery tissue may Celery root tissue was also more inhibitory to lettuce emergence
affect seedling growth more than emergence. than either lamina or petiole tissue. Differences in phytotoxicity
Overall, celery root tissue was more toxic to lettuce growth of various tissues have been reported previously (Guenzi et al.,
than celery petiole and lamina tissue (Table 2). Based on lettuce 1967; May and Ash, 1990; Putnam and Tang, 1986; Rice, 1974).
shoot weight, celery root tissue was 1.8 and 1.6 times more Such differences might be related to allelopathic compounds
toxic than celery petiole and lamina tissue, respectively. Based being produced in larger quantities in certain tissues, imparting
on lettuce root weight, celery root tissue was 1.8 and 1.3 times a higher level of toxicity. Release of phytotoxic compounds
310 J. Amer. Soc. Hort. Sci. 117(2):308-312. 1992.
could also be affected by tissue type. Leaf and petiole tissues Carbon alone had no effect on the growth of lettuce (data not
are covered with cuticle and are more lignified than root tissue. shown). However, the allelopathic effect of celery residue on
Both of these factors could potentially regulate the release of lettuce growth was reduced as carbon concentration increased
allelopathic compounds. (Table 4). Activated carbon partially reversed the inhibitory ef-
Incubation of celery residue in soil. Incubation enhanced the fect of celery residue on lettuce growth, presumably by adsorb-
toxicity of celery residue at 1% (v/v) only after 4 weeks, but ing phytotoxic compounds produced by celery, but total reversal
progressively reduced it at 4% (v/v) (Table 3). At the lower was absent, probably because the charcoal did not adsorb 100%
percentage, partial tissue degradation may have been necessary of the toxic substances.
to release a sufficient amount of the compound(s) responsible Various factors were shown to influence the magnitude of
for inhibition of lettuce growth. At the higher percentage, the lettuce growth inhibition by celery residue, including 1) tissue
critical amount of allelopathic compound(s) could have been type and amount, 2) growth medium used, 3) incubation of the
released more rapidly. There was no inhibition of lettuce growth residue soil mixture, and 4) the presence of activated carbon.
at the 8- and 12-week incubation periods for l% (v/v), probably Collectively, these data support the hypothesis that celery res-
because the toxic compounds had undergone degradation. Both idue has allelopathic potential. The alteration of C : N ratio in
activation (i.e., incubation enhancing phytotoxicity) and deg- soils, leading to rapid assimilation of N by microorganisms, is
radation (i.e., incubation reducing phytotoxicity) have been re- known to reduce early plant growth (Alexander, 1977). If celery
ported previously (Guenzi et al., 1967; May and Ash, 1990; residue inhibited lettuce growth by altering C : N ratio, then
Nair et al., 1990; Patrick, 1971). adding a small amount of residue to a highly organic medium
These data support the contention that celery residue contains should have had no effect. Incubation also would not have en-
and potentially releases organic compound(s) that exert an al- hanced inhibition, and increased fertility would have overcome
lelopathic effect toward lettuce. One percent (v/v) of celery root inhibition. Lastly, if celery residue had inhibited lettuce growth
residue would be equivalent to 0.5% on a weight-to-weight basis indirectly by altering C : N ratio, the addition of activated car-
for the soil used in this experiment. This degree of toxicity bon would have had no effect.
generally was as, or more, toxic than that of plant residues that It is often difficult to conclude from greenhouse experiments
have been reported (Putnam and Tang, 1986; Rice, 1974). whether an allelopathic effect has actually occurred. Ultimately,
Fertility effect. To further substantiate that the potential al- isolation, characterization, and proof that certain compounds are
lelopathic effect of celery residue on lettuce growth was not due the cause of the inhibitory effect is necessary to unequivocally
to changes in C : N ratio [which would alter nutrient concen- prove allelopathy (Putnam and Tang, 1986).
trations through microbial assimilation (Alexander, 1977)], fer-
tility studies were conducted. Celery root residue reduced lettuce Literature Cited
shoot growth by 70% when grown at the maximum fertility level Alexander, M. 1977. Introduction to soil microbiology. 2nd ed. Wiley,
used in these experiments (Fig. 1). In the presence of celery New York. p. 467.
residue, there was only a slight increase in lettuce shoot growth Baskin, J.M., C.J. Ludlow, T.M. Harris, and F.T. Wolfe. 1967. Psor-
with increasing fertility. Fertility did not significantly affect the alen, an inhibitor in the seeds Psoralea subacaulis (Leguminosea).
response of lettuce root growth to celery root residue (Fig. 2). Photochemistry 6:1209.
The lettuce growth response was probably not due to an altered Beier, R. C., G.W. Ivie, E.H. Oertli, and D.L. Holt. 1983. HPLC
C : N ratio caused by celery residue but instead due to the analysis of linear furocoumarins (psoralens) in healthy celery (Apium
graveolens). Food Chem. Toxicology 21:163-165.
allelopathic nature of celery residue.
Beier, R.C. and E.H. Oertli. 1983. Psoralen and other linear furocou-
Activated carbon effect. This study was conducted to deter- marins as phytoalexins in celery. Photochemistry 22:2595–2597.
mine whether the addition of activated carbon, which is known Brown, S.A. 1981. Coumarins, p. 286-287. In: P.K. Stumpf and E.E.
to bind and inactivate many organic compounds, would reduce Corm (eds.). The biochemistry of plants: A comprehensive treatise
or eliminate the allelopathic effect of celery residue on lettuce vol. 7. Secondary plant products. Academic, New York.
growth; however, absorptivity varies with the chemical nature Draper, N.R. and H. Smith. 1981. Applied regression analysis. 2nd
of the compound (Mattson and Mark, 1971). ed. Wiley, New York. p. 47-49.
J. Amer. Soc. Hort. Sci. 117(2):308-312. 1992. 311
Duke, S.0. 1986. Naturally occurring chemical compounds as herbi- azobenzene a microbially transformed allelochemical from 2,3-ben-
cides. Rev. Weed Sci. 2:17-65. zoxazolinone: I. J. Chem. E C OL 16:353-364.
Fischer, N.H. 1986. The function of mono and sesquiterpenes as Patrick, Z.A. 1971. Phytotoxic substances associated with the decom-
plant germination and growth regulators, p. 203-218. In: A.R. position in soil of plant residues. Soil Sci. 111:13-18.
Putnam and C. Tang (eds. ). The Science of allelopathy. Wiley, Patrick, Z. A., T.A. Toussoun, and L.W. Koch. 1964. Effect of crop-
New York. residue decomposition products on plant roots. Annu. Rev. Phyto-
Guenzi, W. D., T.M. McCalla, and F.A. Norstadt. 1967. Presence, and pathology 2:627-292.
persistence of phytotoxic substances in wheat, oat, corn, and sorghum Patrick, Z. A., T.A. Toussoun, and W.C. Snyder. 1963. Phytotoxic
residues. Agron. J. 59:163-165. substances in arable soils associated with decomposition of plant
Hoagland, D.R. and D.I Amen. 1950. The water-culture method for residues, Phytopathology 53: 152–161.
growing plants without soil. Calif. Agr. Expt. Sta., Berkeley. Circ. Putnam, A.R. 1988. Allelochemicals from plants as herbicides. Weed
347. Technol. 2:510-518.
Lord, K. M., H.A.S. Epton, and R.R. Frost. 1988. Virus infection and Putnam, A.R. and C. Tang. 1986. The science of allelopathy. Wiley,
furocoumarins in celery. Plant Pathology 37:385-389. New York.
MacLeod, A.J., G. MacLeod, and G. Subramanian. 1988. Volatile Rice, E.L. 1974. Allelopathy. Academic, New York.
aroma constituents of celery. Photochemistry 27:373-375. Rice, E.L. 1984. Allelopathy. 2nd ed. Academic, Orlando, Fla.
Mattson, J.S. and H.B. Mark. 1971. Activated carbon. Marcel Dekker, Sherf, A.F. and A.A. MacNab. 1986. Vegetable diseases and their
New York. control. p. 157-201. Wiley, New York.
May, F.E. and J.E. Ash. 1990. An assessment of the allelopathic Surico, G., L. Varvaro, and M. Solfrizzo. 1987. Linear furocoumarin
potential of Eucalyptus. Austral. J. Bet. 38:245-254. accumulation in celery plants infected with Erwinia carotovora pv.
Nair, M. G., C.J. Whitenack, and A.R. Putnam. 1990. 2,2’-OXO-1,1’ carotovora. J. Agr. Food Chem. 35:406409.
312 J. Amer. Soc. Hort. Sci. 117(2):308-312. 1992.